Risk measures with Generalized Secant Hyperbolic Dependence

In this paper we propose to model the dependence of multiple time series returns with a multivariate extension of the generalized secant hyperbolic distribution (GSH) using the NORTA (NORmal-to-Anything) approach and the Koehler and Symanowski copula function. The two methodologies permit to generate random vectors with marginals dis- tributed as a GSH distribution and given correlation matrix, which can be used to measure the risk of a portfolio using the Monte Carlo method.